格雷厄姆与多兹维尔访谈（第5期）

   results depart   the current market? More   trage is also a very plank in   trageurs. Second,   these
   specifically, what changes   finance theory. For example,   theories fall short when we
meaningfully from   have you seen recently in   the Modigliani and Miller   test them empirically. For
   the way that investors be-   capital structure invariance   example, most mean vari-
   normal   have, and in your view, does   propositions from the late   ance models assume normal
   the recent behavior that   1950s are based on an arbi-   or log-normal price changes.
distributions. And   you’re seeing reinforce or   trage argument. And of   But we know that the distri-
 historically when   depart from historical   course the most famous   bution of asset price
   norms?   application of arbitrage is   changes is fat-tailed. The
 there have been   the Black-Scholes option-   real-world results depart
   MM: To answer that I want   pricing model.   meaningfully from normal
   the greatest   to take one step back and   distributions. And histori-
   talk the theories of how   The third way to get effi-   cally when there have been
   arbitrage   market efficiency comes   ciency is what is colloquially   the greatest arbitrage op-
   about. We’ll first ask, how   called the “wisdom of   portunities of all—and we
 opportunities of   and why do markets get to   crowds.” More formally, it is   may be in that situation to-
 all—and we may   efficiency? And then from   viewing markets as a   day—arbitrageurs have
   there we can figure out why   “complex adaptive system.”   failed to show up. Research-
be in that situation   they become inefficient.   The simple story is that   ers documented this pretty

   today—   There are three classical   when a diverse group of   Management in 1998. I think
   ways to explain market effi-   agents get together to solve   it very much characterizes
arbitrageurs have   ciency. The first is mean-   a problem, they tend to   today’s market, but only
   variance efficiency, which is   come up with the right solu-   time will tell.
failed to show up.”   what we teach in business   tion even when any individ-
   school. Here,   investors   ual agent doesn’t know very   So let me turn to the
   understand their utility func-   much. Dumb agents and   “wisdom of crowds” argu-
   tions and how to trade off   collective smarts.   ment. For the wisdom of

models in finance are based   The first two approaches,   to have three conditions in
on mean variance efficiency,   mean variance and no arbi-   place. Condition number
including the capital asset   trage, pose some problems.   one is agent diversity—that
pricing model and most of   First, both make assump-   is, the decision rules or ap-
portfolio theory.   tions about the mechanism.   proaches that the individual

different perspectives. Sec-   mid-1800s.  So  for  people
ond, we need a properly-   with   long  memories—or
functioning   aggregation   large databases—this is not
mechanism—some  way  to   unprecedented.   But   for

tion  into  one  place.  Ex-   we’ve  seen  something  of
changes  obviously  do  this   this magnitude in our invest-
effectively.  And  the  third   ment careers.
thing is properly-functioning
incentives,   effectively   re-   G&D: Would you say that

penalties for being wrong.   served—the  types  of  mis-
pricings due to lack of agent
If  you  apply  those  three   diversity  and  so  forth—
conditions to humans, diver-   reinforce   the   investment

it was very bullish, this is   some  degree.  But  that  is
very bearish. And leverage   really  what  Ben  Graham
has played a meaningful role   called  speculation.  There’s
in this as well, as it has in   nothing  wrong  with  it,  as

know  what  is  normal  in   ideal  type  of  an  environ-
markets. But  I  do  believe   ment. These kinds of mis-
that the period we’re going   pricings  are  fundamentally
through  today is not  that   what you’re looking for.
unique.  Obviously  all  of

some  level,  but  in  some   talk about what it is to be a
ways it is not unique. We’ve   contrarian  investor.  I  sug-
seen  elevated  periods  of   gested   there are two es-
volatility  like  we’re  seeing   sential components to this:
today;  we  saw  it  in  the   be prepared to go against

valuation analysis. To play   strongly as I can. I would   G&D: In that framework,
devil’s advocate and ques-   argue that the single most   isn’t it risky to use market
tion the purpose of the be-   common error in the invest-   expectations as a baseline
havioral phenomenon to   ment industry is a failure to   for your own opinion on a
begin with, at what point   distinguish between funda-   stock—particularly in situa-
could you just value an as-   mentals and expectations.   tions where the market
set, and if it’s trading below   And those are two very   might not be pricing effi-
its value then you should   distinct things that you have   ciently? How do you man-
buy it, regardless of what’s   to think about separately.   age that risk?

   MM:   You’re obviously   the price. What happens   think probabilistically. So
   looking for some sort of   typically is when fundamen-   whereas we all have a view,
   evidence of mispricing in   tals are good, people want   or a most-likely outcome, it
   your search strategy. Let me   to buy. And when funda-   is really crucial to consider
  “I would argue   reiterate the obvious: it is   mentals are bad, people   if/then scenarios. What I like
   very difficult for active man-   want to sell. It’s a very hu-   to do is to take a current
  that the single   agers to beat the market.   man reaction. But the goal is   stock price and reverse-
   most common   We know for sure they   to be very explicit in sepa-   engineer the current expec-
   don’t do it in aggregate be-   rating those two things.   tations—which I think is a
   error in the   cause of fees. We also know   reflection of the current
   it’s very difficult for anybody   For example, let’s say you   consensus. That gives me
   investment   to do it on a sustainable   went to the race track and I   some sense of the baseline.
   basis. So this is not a trivial   told you the [most prob-   And I then consider various
industry is a failure   task. So step one is to figure   able] winner of every race   if/then scenarios. This an-
   out your search strategy.   in advance. That’s not really   swers questions like: if
   to distinguish   We look for diversity   the information you want.   things are better—for ex-
   between   breakdowns as the headline,   What matters is not so   ample, revenues are really
   and there can be all sorts of   much which horse is going   going to come in higher or
fundamentals and   indicators of a breakdown.   to win, but what odds are   margins come in better—
   Examples include statistically   priced in. So you’re not just   what does it mean for value?
  expectations.”   cheap stocks, for example   looking at fundamentals—   And likewise, if things come
   high dividend yields and low   how fast the horse can   in worse, what happens to
   price-to-earnings or low   run—you’re looking at rela-   value?

can be changes in manage-   tals and expectations. And   Whenever I think about an
ment—so a new manage-   it’s the difference between   investment I like to think
ment comes in, reallocating   those two things that is so   about the range of possible
capital in a new way—or a   essential. That’s why I’m   value outcomes and the
change in industry struc-   such an advocate for the   associated probabilities of
ture—for instance, the in-   reverse-engineering expec-   those outcomes. The ex-
dustry’s consolidating, or   tations investing approach.   pected value is the product
expanding geographically, or   Often it is easier to say   of those probabilities and
cutting capacity. So there   what is built into the price,   outcomes. Of course, you
could be a host of items   and I think the fundamentals   need to compare that ex-
that might be a source of an   are going to be better or   pected value to the current
edge and encourage you to   worse. It’s more like betting   price. So while people may
investigate it further.   on the over/under in a foot-   look at something and think

in  setting  these  outcomes   revisit   them   periodically.
and probabilities? Is it more   This  allows  you  to  effec-
of  an  overarching  frame-   tively look back on how you
work  that  you  take  into   were  thinking  about  the
account, or do you explicitly   situation  at  a  given  time,
set  out  your  distributions   and keeps you from falling
and expected values?   for  hindsight  bias.  By  the
way,   Warren Buffett and
MM: We set them out ex-   Charlie  Munger  say  very
plicitly. I’ll make one link to   explicitly  that  it’s  about
one  of  the  behavioral  fi-   probabilities and outcomes.

example of how you might   been  a  lack  of  bullishly-
combat that bias.   inclined  investors  recently
and  a  preponderance  of
One thing you can do is go   bears.  What  other  agents
to the options market and   are  missing  from  today’s

swer—certainly   it’s   the   geurs  that  cruise  markets
market’s answer—but what   and—of  course  in  more
we found almost inevitably   normal   conditions—keep
that in doing that exercise,   things  reasonably  efficient.
the  options  market  re-   But  those  arbitrageurs  at
Page 23

Eventually, almost  no one   with Paul Johnson at this
has access to capital.  One   fall’s Security Analysis Sym-
of  the  key  things  missing   posium.
today is an active commu-
nity  of  arbitrageurs.  And
that’s going to take some
time to repair. We’re seeing
a number of new funds step
in—bankruptcy   or   dis-
tressed funds—and they will
eventually move the assets

   By the way, this is nothing   the contextual factors that   level of earning and capital-
   new. Roger Lowenstein’s   might skew those metrics,   ize it. So you’re only paying
   book, When Genius Failed,   like inflation, taxes, compo-   for the steady-state, not for
Books by Michael   starts off with the story of   sition of assets, and so   the present value of growth
Mauboussin:   John Eckstein, an arbitrageur   forth?   opportunities.

Eckstein had put on a clas-   MM: If you adopt a longer-   The third level is calculating
sic arbitrage position with   term perspective—say three   a franchise value, where
leverage. But the position   to five years, which to me is   you’re paying for future
worked against him, and   a pretty reasonable time to   value creation. There are a
because he was leveraged,   consider—and take a prob-   fair number of businesses
he ran out of capital. So he   abilistic view of things, I   out there that you can buy
had to shop his position,   think the preponderance of   today without paying much,
and the guy who took it   evidence suggests that re-   if anything, for franchise
over was John Meriwether,   turns will be satisfactory and   value. Bruce’s approach is
then at Salomon Brothers.   maybe even really good.   very reasonable in all envi-

exact same position as Eck-   use some of the valuation   The second comment re-
stein when Long-Term   techniques   Bruce   lates to volatility. Volatility is
Capital Management’s port-   Greenwald has popularized.   a broad proxy for the cost
folio started to bleed badly.   You can go back to the Gra-   of capital. Clearly credit
By the way, Meriwether,   ham and Dodd roots. Bruce   spreads—both high-grade
and hence Salomon, made a   has articulated value on   and high-yield credit
lot of money on Eckstein’s   three levels. The first is ba-   spreads—have really blown
position as it worked out.   sic old-fashioned asset value,   out. High-yield spreads, for
But unlike Eckstein, Meri-   where you literally go down   example, in the high teens
wether couldn’t find anyone   the balance sheet item by   suggests very elevated ex-
to assume his positions.   item, and mark things to   pectations for risk and re-
What we’re seeing today is   where you think they are   ward. And then you can
an absence of arbitrage.   reasonable today. I think if   look at the VIX, which is the
While the government may   you go through that exer-   one-year implied volatility
try to step in and be the   cise—an intrinsic value   for the S&P 500. The VIX
buyer of last resort, it’d be   based on pure asset value—   was at a ridiculous level,
better to see these proc-   you can find a lot of things   close to 90, in the fall. But
esses work out organically.   that are pretty attractive   now it’s still in the mid-40s.

use to try to find the areas   If that doesn’t get you any-   tent with returns in the high
of the market that are   where, then you step up to   -teens to low-20s. So in
priced inefficiently, such as   second level—earnings   summary you have below
price-to-book or price-to-   power. Earnings power to-   trend earnings and above
earnings ratios. A lot of in-   day is probably below   trend perceived risk. As
vestors, particularly value   trendline. Trendline earn-   both of those normalize
investors, are excited about   ings are probably in the $70   over time, equity markets
the statistical cheapness of   -$80 range for the S&P 500,   can do okay.
stocks today. Do you share   and for 2009 it will be in the   (Continued on page 9)

the streak is to be longer. If
you look  at  a lot of  the   So  here’s  how  you  might
great  investors  more  re-   apply this to investing.  Ask
cently, a lot of them have   a  person  to  create  two
been   pummeled   pretty   portfolios: one they believe
badly.  How do you assess   would be beat the market

luck. With other games like   time.  What I recommend
chess or checkers, can you   people  do  is  rather  than
lose on purpose?  The an-   focus  on  outcomes,  focus
swer is of course yes be-   on  the  investing  process.

consume  more  in  the  fu-
ture.  So stocks should go
up  over  time,  which  of
course  they  have.  That
doesn’t  mean  they  go  up
every year or even  every
ten years. But over the long
haul equities have delivered
real rates of return in the 6-

   for the longs and shorts. It   couple of years. On the   MM: You need to think
   is true that if you are long-   demand side, the govern-   about the type of system
   only you are looking for   ment has taken a lot of ini-   you’re dealing with. Some
   different signs than if you   tiative to try to get financing   systems are going to be
   are also looking for shorts.   costs down.   Obviously,   much more predictable, and
   But there are a whole host   housing affordability should   sticky, than others. Take,
   of other factors that make   also improve from home   for example, the path of
   shorting very difficult.   prices coming down. But   growth in GDP—you don’t
  “Your need to   even with less new home   get fat tails in GDP num-
   G&D: We talked earlier   supply and greater afforda-   bers. The economy doesn’t
 think about the   about mispricings and bear-   bility, it will take some time   grow 70% in one year and
   ish sentiment. Do you see   for housing prices to stabi-   decline 70% the next year.
  type of system   any fundamental signals   lize.   So there are some systems
  you’re dealing   coming out of all the noise   that are a little more sta-
   and volatility in the markets   I also like to look at wealth   tionary. I believe the earn-
with. Some systems   that worry you going for-   effects. The numbers are   ings outlook is something
   ward? For example, are you   almost incomprehensible—   along those lines.
  are going to be   looking at demographic   the net worth of the U.S.
   changes or other fundamen-   consumer is down about $7   Let’s take corporate reve-
   much more   tal indicators that you think   -$8 trillion year-over-year.   nue in the U.S.—that num-
   are a legitimate cause for   If you apply a 5% wealth   ber is going to grow at
 predictable, and   bearishness right now?   effect to that number, you   some number that approxi-
sticky, than others.   have about $400 billion of   mates GDP over time and
   MM: There are clearly   evaporated purchasing   then there is going to be
Take, for example,   many reasons to be con-   power. Fortunately, that’s   some kind of corporate
   cerned in the short term.   being offset in part by lower   profitability   level—
the path of growth   We will likely continue to   gasoline prices. But if we   measured in margins or
   see substantial increases in   see more difficult equity   return on invested capital.
in GDP—you don’t   U.S. unemployment or if   markets and falling home   It turns out that we have
   deflationary pressures con-   prices, the U.S. consumer   been in a steady 20-25 year
get fat tails in GDP   tinue. The one area that I   will continue to be very   trend of rising returns on
   numbers.”   would try to keep a close   strapped. The U.S. con-   capital. In fact, going into
   eye on is the housing mar-   sumer represents something   this market dislocation, re-
   ket. It’s not an exaggeration   like 17-18% of global GDP,   turns on capital were among
   to say that the roll over in   so that’s a really important   the highest they’ve ever
   housing has led to the cur-   engine that’s sputtering in   been. So clearly returns are
   rent set of problems we are   the global economy.   coming down and may even
   facing. Likewise,   finding   go back to where they were
   some footing in the housing   G&D: Nassim Taleb has   in the 1970s. But I suspect
   market may provide an indi-   written a lot about non-   that because of the compo-
   cator as to when we come   normal and non-stationary   sition of corporate Amer-
   out of this downturn.   environments. When you   ica—that the mix between
   Clearly, from a supply-   mentioned that trendline   the manufacturing busi-
   demand perspective, we’ve   earnings for the S&P are   nesses, service businesses,
   seen pretty meaningful   $70-$80, how confident are   and knowledge businesses—
   changes on both sides.   you in that level? How do   I’m more sanguine   and
   From a supply perspective,   you know that the last 20   (Continued on page 11)

in returns.   market.  I think you could
probably move into some of
Those  kinds  of  trends—   these investments with rela-
GDP   growth,   earnings   tively low risk and relatively

those  are  the  things  that   of return that has been very
don’t so much lend them-   attractive  on  a  historical
selves   to   the   extreme   basis.  Even in the fixed in-
events  as  do  asset  price   come  markets,  whenever
changes  or  other  natural   you  see  fixed  income  re-
things  where  that  would   turns that are competitive
come   into   play—   with  long-term  equity  re-

and so forth.  The point I   good time to get involved.
want to make is that there   I’ve  seen  reports  saying
is a continuum of distribu-   what’s  priced  into  high-
tions.   Some  distributions   grade or at least the highest

reading. You had a lot peo-
ple who saw the perform-
ance  of  alternative  asset
management strategies, and
decided  to  model  them-
selves after the Harvard or
Yale   endowment.   They

out a full understanding of
what was going on.   That   strategic buyers.”
led   to   significant   asset
growth,  but  when  things
turned  sour,   investors
turned  tail.  The  estimates
call for something like $300-
$400 billion of net outflows
from  hedge  funds  on  an
asset base of $1.5-1.6 tril-
lion. Clearly, there will be a
hedge fund industry, includ-
ing  quantitative  strategies,
but there has to be some
sort  of  a  healing  process

(Continued from page 11)   intersection between valua-   can point to that is explana-
ago the PE firms had access   tion work and competitive   tory or causal for high or
to financing at very attrac-   analysis. For example, if you   consistent returns. In other
tive rates. Now they can’t   take a competitive strategy   words, the return patterns
get a loan. Of course, there   course, you will learn a lot   are not much different than
will be a time again when PE   of really cool things about   what chance would dictate,
firms will have access to   industry life cycles, five   So there’s an interesting
capital. Quite frankly, we   forces, value chains, disrup-   intersection between strat-
are heading toward a very   tive innovation, and game   egy and finance—we don’t
attractive mergers and ac-   theory—all these things that   know enough about it yet.

my opinion, for both finan-   tal for businesses. Then you   G&D: Final question. It
cial and strategic buyers. I   take a finance course and   wouldn’t be an appropriate
don’t know who will be the   learn the valuation stuff.   interview if we didn’t ask for
first to dip their toes into   But they are treated as dis-   your advice on pursuing a
the water, but we are set-   tinct disciplines. What I   career in investment man-
ting up for a very attractive   would really love to see is   agement, especially in the
M&A market, maybe in the   some way to combine these   current environment.

-2011.   ple, companies with better   MM: It’s clearly a very
   strategic positions should be   tough environment.   The
G&D: Is that something   accorded higher valuations   one thing I would say is that
you can try to capitalize on?   and poor companies should   it’s very important to find an

MM: It is something we try   ket sorts this out pretty   intellectually compatible
to capitalize on. We’re   well, but to me that’s one of   with. It’s important to en-
thinking about, as part of   the really exciting areas   joy and respect the people
our investment considera-   where we can get more   there, but most important is
tion, the probability of   rigorous.   the work they’re doing and

ing takeover candidates.   About a year ago I published   with your personality. Being
We are thinking about   a piece about mean rever-   somewhere with which you
which industries are most   sion and return on invested   are philosophically aligned is
likely to undergo consolida-   capital. It showed a very   really important.   Some
tion. Clearly, financial ser-   simple, economically intui-   people feel more comfort-
vices are still high on the list   tive conclusion, which is   able in short-term, trading
for a lot of obvious reasons.   that high-return companies   oriented organizations.
We are thinking about   tend to see their return on   Others feel comfortable in a
which industries are ripe for   invested capitals   move   long-term, patient type of an
it and which companies are   down toward the cost of   organization. The key is to
positioned well, both as   capital over time. Likewise,   do a really honest self-
sellers and buyers.   for firms that earn a very   assessment and do good job